Closed Sryborg closed 4 months ago
Hi Sryborg
Thank you! This is very good! Could you please make the following changes to the pull request
(1) on lines 384 and 385 you appear to have only changed the identation by 2 char we prefer to only introduce changes which impact the code (2) on line 390 you have a small change to the error message. Do you think this is important enough to change? (3) on line 397 you have changed the spacing "TLSH_STRING_LEN_REQ-2" => "TLSH_STRING_LEN_REQ - 2" We really do not want to have changes like that (4) The Py_XDECREF(asciiStr); look to be the critical changes - though I note that you improved the readibility of the code while doing your bug fix
Please resubmit merge request with changes
Hi jonjoliver, Thankyou for the comments.
(1) Reverted (2) I had added while debuging, when I was testing with random strings(deliberately shorter TLSH, texts etc). But yeah, in the larger scheme of things, it wouldn't matter, Reverted (3) Reverted. (4) 👍🏼 :)
Hi @Sryborg
Thank you. It appears that the memory leak is due to PyUnicode_AsASCIIString() leaking memory unless you do Py_XDECREF As discussed in https://discuss.python.org/t/memory-leak-with-pyunicode-fromstring-and-pydict-clear/39218
Approved.
Providing a fix to the issue #125 Modified the tlshmodule to fix the memory leak.